Anita Zeng Posted January 22, 2020 Share Posted January 22, 2020 hi~ I'm just beginning to learn GWB. I am confused about the boundary condition setting of a single point source pollution for one-dimensional mass transport. Attached are the questions and the setting of X1t simulation parameters. Maybe I guess there are something wrong on the Intervals pane, so it can't be runing. But I don't know how to modify the setting. Could you please help me, thank you. Question of non-reactive one-dimensional solute transport.pptx 1D nonreact-2.x1t Quote Link to comment Share on other sites More sharing options...

Jia Wang Posted January 23, 2020 Share Posted January 23, 2020 Hello Anita, I took a quick look at your input file and here are a couple of suggestions to help you get started. It seems like you are trying to model a continuously leaking point source contamination for 1 year. However, I notice that in your Intervals pane, you are only setting a one day reaction interval for the 'contaminated' fluid to enter your system at the inlet and the remaining time flushing the system with dilute fluid. If the goal is to simulate a continuously leaking source of contamination at the inlet, then you should remove the second reaction interval from your simulation and allow for the contaminated fluid flow into the system for a full year. To quickly remove any reaction interval period in your Intervals pane, you can simply click to select it and then click the delete button. I deleted the reaction interval that set the inlet fluid as the dilute fluid (named 'fluid 2') and that allow the model to run to completion. For more information on setting boundary conditions and reaction intervals, please refer to section 2.10 and 2.9 in the Reactive Transport User's Guide. You might also be interested in looking at X1t examples listed in section 3.5. The input file corresponding to each example in section 3.5 is available in the GWB installation subdirectory 'Script'. Hope this helps, Jia Wang Quote Link to comment Share on other sites More sharing options...

Anita Zeng Posted February 6, 2020 Author Share Posted February 6, 2020 On 1/24/2020 at 7:37 AM, Jia Wang said: Hello Anita, I took a quick look at your input file and here are a couple of suggestions to help you get started. It seems like you are trying to model a continuously leaking point source contamination for 1 year. However, I notice that in your Intervals pane, you are only setting a one day reaction interval for the 'contaminated' fluid to enter your system at the inlet and the remaining time flushing the system with dilute fluid. If the goal is to simulate a continuously leaking source of contamination at the inlet, then you should remove the second reaction interval from your simulation and allow for the contaminated fluid flow into the system for a full year. To quickly remove any reaction interval period in your Intervals pane, you can simply click to select it and then click the delete button. I deleted the reaction interval that set the inlet fluid as the dilute fluid (named 'fluid 2') and that allow the model to run to completion. For more information on setting boundary conditions and reaction intervals, please refer to section 2.10 and 2.9 in the Reactive Transport User's Guide. You might also be interested in looking at X1t examples listed in section 3.5. The input file corresponding to each example in section 3.5 is available in the GWB installation subdirectory 'Script'. Hope this helps, Jia Wang Dear Jia ~ Thanks for your response. In this problem, Cl pollutant is instantaneous leaking to groundwater layer, then compute the Ccl- at d=15m after 1y from the leaking point. Therefore, I set two different fluids in Fluids pane and two reaction interval in Intervals pane. If I set one minute(or one second) reaction interval for the contaminated fluid to enter the system at the inlet and the remaining time flushing the system with dilute fluid, the program cat't run. It shows "residuals too large, 669-th interation, Newton failed at node 0, ............ , Abandoning simulation". What's wrong with this. How can I correct the setting so that the program can run normally? Thank you. 1D nonreact-6instant.x1t Quote Link to comment Share on other sites More sharing options...

Jia Wang Posted February 14, 2020 Share Posted February 14, 2020 Hello Anita, If you set up your system to have the contaminant flowing into the domain for an interval period of 1 minute, the concentration of Cl- you will see in the system is going to be much lowered than your expected results. To check, you can set the endpoint of your simulation to 10 minutes (before the simulation fails) and see that the maximum concentration of Cl- by plotting the results. You can plot X-position (on X Axis) vs. Cl- in fluid (mg/l) (on Y Axis) and cycle through the time level in the XY configuration dialog. You will see that the concentration of Cl- in your system is only ever as high as 2.18 mg/l. If you are expecting to see a pulse of contaminant in your domain with the Cl- concentration of 725 mg/l, you will need to have a longer interval where the contaminant is flowing into the system. I tried increasing the contaminant interval to 40 days and was able to arrive at 30 mg/l of Cl- at 15 meters from the leaking point. The error you are seeing are associated with stability issues with in your model. You can try to setting your dx_init, initial time step, to a small value (like 1e-5) to help with stability. To change dx_init, go to 'Config' -> 'Stepping' and you will see an option to set a dx_init value. When I tried using 1e-5 for the initial time step, the model ran to completion but again, I am seeing much lower concentration at 15 m than the answer you expected. Can you perhaps clarify what you mean by instantaneous leaking of the contaminant? And is the 30 mg/l of Cl- an analytical solution? Best regards, Jia Wang Quote Link to comment Share on other sites More sharing options...

Anita Zeng Posted February 15, 2020 Author Share Posted February 15, 2020 Dear Jia ~ Thanks again for your assistance and teaching. The example is from a textbook about mass transport of solutes. You can see the attached file for the original question. Cl- 30mg/L is the value of analytical solution calculated by the formula. I want to get the value of numerical solution by GWB, then confirm the difference between these two method. If I model a continuously leaking point source contamination for 1 year, the wave crest of solute don’t move in the resulting graph (X-position vs. Cl- in fluid (mg/l)) with time and distance. So I think it’s not continuously leaking for 1 year. The time for contaminant leaking should be very short, so we can see the moving of solute slug with time and distance (like figure 11.11 in attached file). Thank you. example.pdf Quote Link to comment Share on other sites More sharing options...

Jia Wang Posted February 19, 2020 Share Posted February 19, 2020 Hello Anita, Thank you for the additional information. In Ogata 1970, it was stated that the concentration at x = 0 is maintained at concentration C0 through time to arrive at the analytical solution used in your textbook. Please see example 1 and 2 in the paper for details. Furthermore, the problem prompt from your textbook does not state when the leachate stopped leaking into your system. To accurately treat a solute slug problem, like you described above, you would need that information. Therefore, I think you should approach this problem as if the leachate is leaking continuously for the 1-year duration that you are modeling. A couple of other suggestions that might helpful here. You might want to consider extending the length of your domain by a small distance, maybe by 5 or 10 meters. Dispersion is not evaluated at the boundary condition and therefore introduce a bit of error to the node at the boundary. I would suggest you extend your domain length and look at the concentration at the nearest node to 15 meters. It might also be helpful to increase the number of nodes in your domain. With numerical modeling, there's a certain amount of numerical dispersion associated with grid sizes. Increasing the number of nodes would generally help reduce that numerical dispersion. In your case, I doubled the number of nodes to 300 and it helped. However, at the end of the day, model results will differ slightly from your analytical solution due to the estimations made in the numerical method and also consideration of additional variables that the analytical solution might not. For example, the GWB also considers variables like density changes with varying solute concentration which is not considered in the analytical solution. Hope this helps, Jia Wang Quote Link to comment Share on other sites More sharing options...

## Recommended Posts

## Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.